Semantic Mapping Based on Ontology and a Bayesian Network and Its Application to CAD and PDM Integration

In a Collaborative Product Commerce (CPC) environment, it is necessary that the participants in a product life cycle should share semantics of terms although they may be represented differently. In order to manage this sharing of semantics, it is necessary to recognize automatically that two terms represented differently can have equivalent semantics. To this end, a semantic mapping logic that utilizes ontology and a Bayesian Network is proposed. The proposed approach consists of three phases: character matching, definition comparisons and similarity checking. First, character matching maps two terms that have identical character strings; second, the definition comparison step compares the two terms using their ontological definitions. Finally, similarity checking evaluates the similarity between two terms using their ontological structure and the Bayesian network. This final phase consists of three steps. Firstly, it calculates similarity between two terms in terms of their character strings and ontological definitions. After this step, it constructs a Bayesian network with the paired terms based on their ontological structure. Finally, it infers whether the pairs are mapped based on the network through a probability equation. The proposed approach is also applied to the integration of the CAD and PDM systems.